US10069691B2 - Method and apparatus for anomaly detection in a network - Google Patents

Method and apparatus for anomaly detection in a network Download PDF

Info

Publication number
US10069691B2
US10069691B2 US15/039,064 US201315039064A US10069691B2 US 10069691 B2 US10069691 B2 US 10069691B2 US 201315039064 A US201315039064 A US 201315039064A US 10069691 B2 US10069691 B2 US 10069691B2
Authority
US
United States
Prior art keywords
sub
stream
network
data
phase space
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related, expires
Application number
US15/039,064
Other languages
English (en)
Other versions
US20170026250A1 (en
Inventor
Tracy Hu
Qingyan LIU
Sky Zhao
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Telefonaktiebolaget LM Ericsson AB
Original Assignee
Telefonaktiebolaget LM Ericsson AB
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Telefonaktiebolaget LM Ericsson AB filed Critical Telefonaktiebolaget LM Ericsson AB
Assigned to TELEFONAKTIEBOLAGET LM ERICSSON (PUBL) reassignment TELEFONAKTIEBOLAGET LM ERICSSON (PUBL) ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: LIU, Qingyan, HU, TRACY, ZHAO, Sky
Publication of US20170026250A1 publication Critical patent/US20170026250A1/en
Application granted granted Critical
Publication of US10069691B2 publication Critical patent/US10069691B2/en
Expired - Fee Related legal-status Critical Current
Adjusted expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/142—Network analysis or design using statistical or mathematical methods
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/147—Network analysis or design for predicting network behaviour
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
    • H04L41/5003—Managing SLA; Interaction between SLA and QoS
    • H04L41/5009—Determining service level performance parameters or violations of service level contracts, e.g. violations of agreed response time or mean time between failures [MTBF]

Definitions

  • the disclosure generally relates to network performance monitoring, and more particularly, to a method and apparatus for anomaly detection in a network as well as associated computer program and computer program product.
  • KPI data is used as a measure of network performance or service quality provided by a network.
  • the KPI data is known as a category of time-series data.
  • off-line statistical analysis methods are applied to time-series data, i.e., the time-series data needs to be stored first and then compared with one or more defined thresholds, so as to find an anomaly.
  • many manual operations are required to define and adjust the thresholds.
  • such manual operations become impractical as the number of network elements increases. It is thus a challenge to detect anomalies in the KPI data in a timely manner.
  • WO 2009/008783A1 discloses a method for measuring a performance indicator with reduced complexity.
  • a performance filter is applied to a raw data and the performance indicator is calculated from the filtered data.
  • the performance indicator is calculated from the filtered data.
  • a method for anomaly detection in a network comprises: obtaining a stream of time-series data related to the network; and dividing the stream into a number of sub-streams each corresponding to a category of data.
  • the method further comprises, for each of the sub-streams: reconstructing a plurality of phase spaces; predicting, in each of the plurality of phase spaces, whether a data item in the sub-stream is an anomaly candidate based on a prediction model associated with the phase space; and detecting the data item as an anomaly when it is predicted as an anomaly candidate in all of the plurality of phase spaces.
  • each of the plurality of phase spaces is represented by feature vectors each associated with a data item in the sub-stream.
  • the above predicting step comprises: predicting, in each of the plurality of phase spaces, a data item in the sub-stream as an anomaly candidate when its associated feature vector is determined as abnormal based on the prediction model associated with the phase space.
  • the prediction model represents a predetermined range in its associated phase space and any feature vector outside the predetermined range is determined as abnormal.
  • the prediction model is initially constructed from a training data set and periodically updated based on data items that are not detected as anomalies.
  • the prediction model is based on One Class Support Vector Machine (OCSVM).
  • OCSVM One Class Support Vector Machine
  • each of the plurality of phase spaces is reconstructed by applying an embedding function having a pair of dimension and lag as its arguments.
  • the above steps of reconstructing, predicting and detecting are performed in a distributed manner over more than one computing node.
  • the time-series data comprises Key Performance Indicator (KPI) data which is a measure of network performance or service quality provided by the network.
  • KPI Key Performance Indicator
  • the method further comprises a step of reporting the data item that is detected as an anomaly to a computer device for providing information about the anomaly.
  • an apparatus for anomaly detection in a network comprises: a unit for obtaining a stream of time-series data related to the network; a unit for dividing the stream into a number of sub-streams each corresponding to a category of data; a unit for reconstructing, for each of the sub-streams, a plurality of phase spaces; a unit for predicting, in each of the plurality of phase spaces constructed for each of the sub-streams, whether a data item in the sub-stream is an anomaly candidate based on a prediction model associated with the phase space; and a unit for detecting, for each of the sub-streams, the data item as an anomaly when it is predicted as an anomaly candidate in all of the plurality of phase spaces.
  • each of the plurality of phase spaces is represented by feature vectors each associated with a data item in the sub-stream.
  • the unit for predicting is configured to predict, in each of the plurality of phase spaces, a data item in the sub-stream as an anomaly candidate when its associated feature vector is determined as abnormal based on the prediction model associated with the phase space.
  • the prediction model represents a predetermined range in its associated phase space and any feature vector outside the predetermined range is determined as abnormal.
  • the prediction model is initially constructed from a training data set and periodically updated based on data items that are not detected as anomalies.
  • the prediction model is based on One Class Support Vector Machine (OCSVM).
  • OCSVM One Class Support Vector Machine
  • each of the plurality of phase spaces is reconstructed by applying an embedding function having a pair of dimension and lag as its arguments.
  • the time-series data comprises Key Performance Indicator (KPI) data which is a measure of network performance or service quality.
  • KPI Key Performance Indicator
  • the apparatus further comprises: a unit for reporting the data item that is detected as an anomaly to a computer device for providing information about the anomaly.
  • an apparatus comprising a processor and a memory.
  • the memory comprises instructions executable by the processor whereby the apparatus is operative to obtain a stream of time-series data related to a network; and divide the stream into a number of sub-streams each corresponding to a category of data.
  • the memory further contains instructions executable by the processor whereby the apparatus is operative to, for each of the sub-streams: reconstruct a plurality of phase spaces; predict, in each of the plurality of phase spaces, whether a data item in the sub-stream is an anomaly candidate based on a prediction model associated with the phase space; and detect the data item as an anomaly when it is predicted as an anomaly candidate in all of the plurality of phase spaces.
  • each of the plurality of phase spaces is represented by feature vectors each associated with a data item in the sub-stream.
  • Said predicting comprises predicting, in each of the plurality of phase spaces, a data item in the sub-stream as an anomaly candidate when its associated feature vector is determined as abnormal based on the prediction model associated with the phase space.
  • the prediction model represents a predetermined range in its associated phase space and any feature vector outside the predetermined range is determined as abnormal.
  • the prediction model is initially constructed from a training data set and periodically updated based on data items that are not detected as anomalies.
  • the prediction model is based on One Class Support Vector Machine (OCSVM).
  • OCSVM One Class Support Vector Machine
  • each of the plurality of phase spaces is reconstructed by applying an embedding function having a pair of dimension and lag as its arguments.
  • the time-series data comprises Key Performance Indicator (KPI) data which is a measure of network performance or service quality.
  • KPI Key Performance Indicator
  • said memory further contains instructions executable by said processor whereby said apparatus is operative to report the data item that is detected as an anomaly to a computer device for providing information about the anomaly.
  • a computer program comprises computer readable instructions which, when run on an apparatus, cause the apparatus to: obtain a stream of time-series data related to a network; divide the stream into a number of sub-streams each corresponding to a category of data.
  • the computer program further comprises computer readable instructions which, when run on the apparatus, cause the apparatus to, for each of the sub-streams: reconstruct a plurality of phase spaces; predict, in each of the plurality of phase spaces, whether a data item in the sub-stream is an anomaly candidate based on a prediction model associated with the phase space; and detect the data item as an anomaly when it is predicted as an anomaly candidate in all of the plurality of phase spaces.
  • a computer program storage product comprises computer readable storage means storing the computer program of the above fourth aspect.
  • a distributed computing system for anomaly detection in a network comprises a plurality of computing nodes, at least one of which is configured to: obtain a stream of time-series data related to the network; divide the stream into a number of sub-streams each corresponding to a category of data; and distribute each of the sub-streams to a first computing node.
  • Each first computing node is configured to: reconstruct, for the sub-stream distributed to the first computing node, a plurality of phase spaces and distribute each of the plurality of phase spaces to a second computing node for predicting whether a data item in the sub-stream is an anomaly candidate based on a prediction model associated with the phase space; and detect the data item as an anomaly when it is predicted as an anomaly candidate in all of the plurality of phase spaces.
  • the time-series data related to the network is processed as a data stream.
  • the stream is divided into sub-streams each corresponding to a category of data.
  • more than one phase space is reconstructed.
  • it is detected whether a data item in the sub-stream is an anomaly candidate based on a prediction model associated with the phase space.
  • a data item is detected as an anomaly when it is predicted as an anomaly candidate in all of the phase spaces.
  • prediction model eliminates the need for manual operations which would be necessary in threshold-based solutions. In this way, it is possible to detect anomalies from a large amount of network data in a more timely manner. Furthermore, for each category of data, a number of phase spaces are reconstructed and a data item is determined as an anomaly only when it is predicted as an anomaly candidate in all of the phase spaces, which allows a comprehensive analysis of the data item.
  • FIG. 1 is a flowchart illustrating a method for anomaly detection in a network according to an embodiment of the disclosure
  • FIG. 2 is a schematic diagram of a prediction model according to an embodiment of the disclosure.
  • FIG. 3 is a schematic diagram showing a distributed computing scenario according to an embodiment of the disclosure.
  • FIG. 4 is a schematic diagram of an apparatus for anomaly detection according to an embodiment of the disclosure.
  • FIG. 5 is a schematic diagram of an arrangement that can be used in an apparatus for anomaly detection according to an embodiment of the disclosure.
  • FIG. 1 is a flowchart illustrating a method 100 for anomaly detection in a network according to an embodiment of the disclosure. The method includes the steps as described below.
  • a stream of time-series data related to the network is obtained, e.g., by receiving the stream from a base station via a socket connection.
  • the time-series data can be the Key Performance Indicator (KPI) data as mentioned above. It is to be noted here that the disclosure is not limited to the KPI data, but is applicable to other time-series data representing network performance or service quality provided by the network.
  • KPI Key Performance Indicator
  • the stream is divided into a number of sub-streams each corresponding to a category of data.
  • KPI data there are several categories of KPI data, including Base Station Controller Answer Input/Output, Base Station Controller Answer Rate Input/Output, Base Station Controller Location Update Success Rate and Base Station Controller Paging Success Rate.
  • Each KPI data item in the stream may have a KPI category identifier identifying the category to which it belongs.
  • a KPI data stream can be divided into several sub-streams based on the KPI category identifiers of each KPI data item in the stream.
  • Each sub-stream may contain KPI data items having the same KPI category identifier.
  • the subsequent steps S 130 -S 150 are carried out.
  • the steps S 130 - 150 will be given with respect to only one of the sub-streams, denoted as SS, it also applies to any other sub-stream resulted from the step S 120 .
  • each of the phase spaces can be reconstructed by applying an embedding function having a pair of dimension and lag as its arguments.
  • the embedding function, F can be represented as:
  • ⁇ (•) denotes a map function that transforms its input from an original space to a feature space
  • [•] T denotes transposition
  • y(t) denotes a feature vector in the resulted phase space.
  • each of the phase spaces reconstructed in the step S 130 can be represented by feature vectors each associated with a data item in the sub-stream SS.
  • each phase space is composed of feature vectors representing the data items in the sub-stream SS with a particular pair of dimension and lag.
  • step S 140 in each of the plurality of phase spaces reconstructed for the sub-stream SS, it is predicted whether a data item in the sub-stream SS is an anomaly candidate based on a prediction model associated with the phase space.
  • a prediction model associated with the phase space For the purpose of illustration, the following description of the step S 140 will be given with respect to only one of the phase spaces, denoted as PS. However, as can be appreciated by those skilled in the art, it also applies to any other phase spaces resulted from the step S 130 .
  • a data item in the sub-stream SS is predicted as an anomaly candidate when its associated feature vector is determined as abnormal based on the prediction model associated with the phase space PS.
  • FIG. 2 is a schematic diagram of an exemplary prediction model in the phase space PS.
  • the prediction model shown in FIG. 2 is based on One Class Support Vector Machine (OCSVM). It can be appreciated by those skilled in the art that the disclosure is not limited to OCSVM and other model-based prediction techniques, such as Autoregressive Integrated Moving Average Model (ARIMA), can also be used.
  • OCSVM One Class Support Vector Machine
  • ARIMA Autoregressive Integrated Moving Average Model
  • each feature vector in the phase space PS is mapped onto a dot in the two-dimensional x-y plane.
  • the prediction model can also be represented in a three or even higher dimensional space, depending on the dimension M of the embedding function used for reconstructing the phase space PS.
  • the prediction model represents a predetermined range in its associated phase space PS (also referred to as “normal range” hereafter) and any feature vector outside the predetermined range is determined as abnormal.
  • the region labeled “+1” in FIG. 2 denotes the normal range.
  • the feature vectors represented by the dots within the “+1” region are considered to be “normal”; whereas the other vectors represented by the dots outside the “+1” region (i.e., within the region labeled “ ⁇ 1” in FIG. 2 ) are determined as abnormal.
  • the prediction model is initially constructed from a training data set containing data samples that are determined to be normal by e.g., a network operator or a service provider (for example, by means of threshold-based comparison). Then, it can be periodically updated based on data items that are not detected as anomalies. More specifically, the feature vectors associated with the data items that are not detected as anomalies can be used as the basis for updating the prediction model. In this way, the prediction model has a self-learning capability and can be automatically updated without any human intervention.
  • a data item in the sub-stream SS is detected as an anomaly when it is predicted as an anomaly candidate in all of the plurality of phase spaces reconstructed for the sub-stream SS.
  • the data item detected as an anomaly can be reported to a computer device for providing information about the anomaly.
  • the computer device can store and provide this information to a network administrator such that the administrator can take appropriate measures to solve problems associated with the anomaly.
  • the computer device can provide a warning message about the anomaly to the network administrator and then the network administrator may confirm and fix the associated problem by e.g., adjusting an antenna in the base station controller.
  • the number of phase spaces reconstructed for one category of data can be independent from the number of phase spaces reconstructed for any other category of data, which provides an improved flexibility in detecting anomalies from different categories of data.
  • the division of the stream into sub-streams provides potential for distributed processing.
  • the above steps S 130 -S 150 can be performed in a distributed manner over more than one computing node.
  • the distributed computing system 300 includes a plurality of computing nodes. At least one of these computing nodes, e.g., a data collection and distribution node 310 , obtains a stream of time-series data related to the network, divides the stream into a number of sub-streams each corresponding to a category of data, and distributes each of the sub-streams to a phase space reconstruction node 320 - 1 , . . . , 320 - n .
  • the phase space reconstruction node 320 - 1 For the purpose of illustration, the following description will be given with respect to the phase space reconstruction node 320 - 1 . However, as can be appreciated by those skilled in the art, it also applies to any other phase space reconstruction nodes 320 - 2 , . . . , 320 - n.
  • the phase space reconstruction node 320 - 1 reconstructs, for the sub-stream distributed to that phase space reconstruction node, a number, K 1 , of phase spaces and distributes each of the K 1 phase spaces to a prediction node 330 - 1 -K 1 , . . . , 330 - 1 -Kn for predicting whether a data item in the sub-stream is an anomaly candidate based on a prediction model associated with the phase space.
  • the prediction results from the prediction nodes 330 - 1 -K 1 , . . . , 330 - 1 -Kn are output to the detection node 340 - 1 , which detects the data item as an anomaly when it is predicted as an anomaly candidate in all of the plurality of phase spaces.
  • the phase space reconstruction node 320 - 1 can also serve as the detection node 340 - 1 .
  • the prediction results from the prediction nodes 330 - 1 -K 1 , . . . , 330 - 1 -Kn can be fed back to the phase space reconstruction node 320 - 1 for anomaly detection.
  • each of the computing nodes shown in FIG. 3 can be implemented as a Processing Element (PE) in a stream processing framework such as S4, Storm, etc.
  • PE Processing Element
  • the distributed computing system 300 can be implemented based on real-time Hadoop frameworks such as Impala and Spark.
  • FIG. 4 is a schematic diagram of an apparatus 400 for anomaly detection in a network according to an embodiment of the disclosure.
  • the apparatus 400 includes a communication unit 410 for communicating with other entities such as a data source and a computer device for providing information about any detected anomaly.
  • the apparatus 400 further includes an arrangement 420 for implementing the method described above with reference to FIG. 1 .
  • the apparatus 400 may further comprise one or more memories 330 and one or more further functional units 340 , such as a functional unit for generating an anomaly detection report.
  • the arrangement 420 can be implemented as a pure hardware solution or as a combination of software and hardware, e.g., by one or more of: a processor or a micro processor and adequate software and memory for storing of the software, a Programmable Logic Device (PLD) or other electronic component(s) or processing circuitry configured to perform the actions described above, and illustrated, e.g., in FIG. 3 .
  • the arrangement 420 may be implemented and/or described as follows
  • the apparatus 400 includes an obtaining unit 421 that obtains a stream of time-series data related to the network.
  • the apparatus 400 further includes a dividing unit 422 that divides the stream into a number of sub-streams each corresponding to a category of data.
  • the apparatus 400 further includes a reconstructing unit 423 that reconstructs, for each of the sub-streams, a plurality of phase spaces.
  • the apparatus 400 further includes a predicting unit 424 that predicts, in each of the plurality of phase spaces constructed for each of the sub-streams, whether a data item in the sub-stream is an anomaly candidate based on a prediction model associated with the phase space.
  • the apparatus 400 further includes a detecting unit 425 that detects, for each of the sub-streams, the data item as an anomaly when it is predicted as an anomaly candidate in all of the plurality of phase spaces.
  • each of the plurality of phase spaces is represented by feature vectors each associated with a data item in the sub-stream and the predicting unit 424 is configured to predict, in each of the plurality of phase spaces, a data item in the sub-stream as an anomaly candidate when its associated feature vector is determined as abnormal based on the prediction model associated with the phase space.
  • the prediction model represents a predetermined range in its associated phase space and any feature vector outside the predetermined range is determined as abnormal.
  • the prediction model is initially constructed from a training data set and periodically updated based on data items that are not detected as anomalies.
  • the prediction model is based on One Class Support Vector Machine (OCSVM).
  • OCSVM One Class Support Vector Machine
  • each of the plurality of phase spaces is reconstructed by applying an embedding function having a pair of dimension and lag as its arguments.
  • the time-series data comprises Key Performance Indicator (KPI) data which is a measure of network performance or service quality.
  • KPI Key Performance Indicator
  • the apparatus 400 further includes a reporting unit (not shown) that reports the data item that is detected as an anomaly to a computer device for providing information about the anomaly
  • FIG. 5 shows an embodiment of an arrangement 500 which may be used in the apparatus 400 .
  • the arrangement 500 includes a processor 510 , e.g., a Digital Signal Processor (DSP).
  • the processor 510 may be a single unit or a plurality of units to perform different actions of procedures described herein.
  • the arrangement 500 may also comprise an input/output unit 530 for receiving/sending signal from/to other entities.
  • DSP Digital Signal Processor
  • the arrangement 500 includes at least one computer program storage product 520 in the form of a non-volatile or volatile memory, e.g., an Electrically Erasable Programmable Read-Only Memory (EEPROM), a flash memory and a hard drive.
  • EEPROM Electrically Erasable Programmable Read-Only Memory
  • the computer program storage product 520 comprises or is, irrespective of embodiment, a non-transitory computer readable medium.
  • the computer program storage product 520 includes a computer program 540 , which includes code/computer readable instructions, which when executed by the processor 510 in the arrangement 500 causes the arrangement 500 and/or the apparatus 400 in which it is included to perform the actions, e.g., of the procedure described earlier in conjunction with FIG. 1 .
  • the computer program 540 may be configured as a computer program code comprising computer program parts 540 A- 540 D.
  • the code in the computer program of the arrangement 500 includes an obtaining part 540 A for obtaining a stream of time-series data related to a network.
  • the code in the computer program of the arrangement 500 further includes a dividing part 540 B for dividing the stream into a number of sub-streams each corresponding to a category of data.
  • the code in the computer program of the arrangement 500 further comprises a reconstructing part 540 C for reconstructing, for each of the sub-streams, a plurality of phase spaces.
  • the code in the computer program of the arrangement 500 further includes a predicting part 540 D for predicting, in each of the plurality of phase spaces constructed for each of the sub-streams, whether a data item in the sub-stream is an anomaly candidate based on a prediction model associated with the phase space.
  • the code in the computer program of the arrangement 500 further comprises a detecting part 540 E for detecting, for each of the sub-streams, the data item as an anomaly when it is predicted as an anomaly candidate in all of the plurality of phase spaces.
  • the computer program parts could essentially perform the actions of the flow illustrated in FIG. 1 , to emulate the arrangement 420 in the apparatus 400 .
  • the different computer program parts when executed in the processor 510 , they may correspond, e.g., to the units 421 - 424 of FIG. 4 .
  • code means in the embodiments disclosed above in conjunction with FIG. 5 are implemented as computer program modules which when executed in the processing unit causes the device to perform the actions described above in conjunction with the figures mentioned above, at least one of the code means may in alternative embodiments be implemented at least partly as hardware circuits.
  • the processor may be a single CPU (Central processing unit), but could also comprise two or more processing units.
  • the processor may include general purpose microprocessors; instruction set processors and/or related chips sets and/or special purpose microprocessors such as Application Specific Integrated Circuit (ASICs).
  • ASICs Application Specific Integrated Circuit
  • the processor may also comprise board memory for caching purposes.
  • the computer program may be carried by a computer program product connected to the processor.
  • the computer program product may comprise a computer readable medium on which the computer program is stored.
  • the computer program product may be a flash memory, a Random-access memory (RAM), a Read-Only Memory (ROM), or an EEPROM, and the computer program modules described above could in alternative embodiments be distributed on different computer program products in the form of memories.

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Physics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Pure & Applied Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Algebra (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Debugging And Monitoring (AREA)
US15/039,064 2013-11-26 2013-11-26 Method and apparatus for anomaly detection in a network Expired - Fee Related US10069691B2 (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2013/087844 WO2015077917A1 (fr) 2013-11-26 2013-11-26 Procédé et appareil de détection d'anomalie dans un réseau

Publications (2)

Publication Number Publication Date
US20170026250A1 US20170026250A1 (en) 2017-01-26
US10069691B2 true US10069691B2 (en) 2018-09-04

Family

ID=53198171

Family Applications (1)

Application Number Title Priority Date Filing Date
US15/039,064 Expired - Fee Related US10069691B2 (en) 2013-11-26 2013-11-26 Method and apparatus for anomaly detection in a network

Country Status (4)

Country Link
US (1) US10069691B2 (fr)
EP (1) EP3075102B1 (fr)
CN (1) CN105745868B (fr)
WO (1) WO2015077917A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10548032B2 (en) * 2018-01-26 2020-01-28 Verizon Patent And Licensing Inc. Network anomaly detection and network performance status determination

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
NL2015680B1 (en) * 2015-10-29 2017-05-31 Opt/Net Consulting B V Anomaly detection in a data stream.
CN106685674B (zh) * 2015-11-05 2020-01-10 华为技术有限公司 网络事件预测以及建立网络事件预测模型的方法和装置
CN108173670B (zh) * 2016-12-07 2020-06-02 华为技术有限公司 检测网络的方法和装置
CN108228442B (zh) * 2016-12-14 2020-10-27 华为技术有限公司 一种异常节点的检测方法及装置
WO2018133924A1 (fr) * 2017-01-17 2018-07-26 Telefonaktiebolaget Lm Ericsson (Publ) Procédés et appareil d'analyse des performances d'un réseau de télécommunications
US10346756B2 (en) * 2017-01-25 2019-07-09 Centurylink Intellectual Property Llc Machine discovery and rapid agglomeration of similar states
US10438124B2 (en) 2017-01-25 2019-10-08 Centurylink Intellectual Property Llc Machine discovery of aberrant operating states
US20190392350A1 (en) * 2017-02-02 2019-12-26 Nippon Telegraph And Telephone Corporation Feature value generation device, feature value generation method, and program
CN108926352B (zh) * 2017-05-22 2020-10-09 北京大学 一种驾驶疲劳检测方法及系统
CN109947760A (zh) * 2017-07-26 2019-06-28 华为技术有限公司 一种挖掘kpi根因的方法及装置
CN110858195A (zh) * 2018-08-17 2020-03-03 空气磁体公司 时间系列度量的高效的存储和查询
CN110139315B (zh) * 2019-04-26 2021-09-28 东南大学 一种基于自我学习的无线网络故障检测方法
US11574461B2 (en) * 2020-04-07 2023-02-07 Nec Corporation Time-series based analytics using video streams
CN114286365B (zh) * 2021-12-16 2023-09-26 四川通信科研规划设计有限责任公司 一种基于基站电调天线aisg的数据回传方法及系统

Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040039968A1 (en) * 2000-09-11 2004-02-26 Kimmo Hatonen System, device and method for automatic anomally detection
US20080310735A1 (en) * 2007-06-15 2008-12-18 Los Alamos National Security Llc Anomalous change detection in imagery
WO2009008783A1 (fr) 2007-07-11 2009-01-15 Telefonaktiebolaget Lm Ericsson (Publ) Procédé et appareil pour déterminer une performance de service
US20090044065A1 (en) * 2005-03-11 2009-02-12 Matsushita Electric Industrial Co., Ltd. Retransmitting method, radio receiving apparatus, and multiantenna radio communication system
US20090112932A1 (en) 2007-10-26 2009-04-30 Microsoft Corporation Visualizing key performance indicators for model-based applications
US20090238255A1 (en) * 2008-03-24 2009-09-24 Hong Kong Applied Science And Technology Research Institute Co, Ltd. Estimation of Error Propagation Probability to Improve Performance of Decision-Feedback Based Systems
US20100027432A1 (en) 2008-07-31 2010-02-04 Mazu Networks, Inc. Impact Scoring and Reducing False Positives
CN101800674A (zh) 2010-02-21 2010-08-11 浪潮通信信息系统有限公司 基于分流定向的旁路式流量检测模型
CN101854404A (zh) 2010-06-04 2010-10-06 中国科学院计算机网络信息中心 检测域名系统异常的方法和装置
US20100299287A1 (en) 2009-05-22 2010-11-25 Alcatel-Lucent Usa Inc. Monitoring time-varying network streams using state-space models
US20110191076A1 (en) * 2008-10-09 2011-08-04 Shunji Maeda Error detection method and system
US8254391B2 (en) * 2009-04-04 2012-08-28 Oracle International Corporation Method and system for performing blocking of messages on errors in message stream
US20130067106A1 (en) 2011-09-10 2013-03-14 Choudur Lakshminarayan Multi-regime detection in streaming data
US20140195184A1 (en) * 2011-08-17 2014-07-10 Hitachi, Ltd Anomaly Detection/Diagnostic Method and Anomaly Detection/Diagnostic System
US9019839B2 (en) * 2012-08-19 2015-04-28 Audiocodes Ltd. Device, system, and method of monitoring fax relay

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101753381B (zh) * 2009-12-25 2012-10-10 华中科技大学 一种检测网络攻击行为的方法
CN102004859B (zh) * 2010-12-01 2013-06-05 南京工业大学 建筑异常能耗支路在线诊断方法

Patent Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040039968A1 (en) * 2000-09-11 2004-02-26 Kimmo Hatonen System, device and method for automatic anomally detection
US20090044065A1 (en) * 2005-03-11 2009-02-12 Matsushita Electric Industrial Co., Ltd. Retransmitting method, radio receiving apparatus, and multiantenna radio communication system
US20080310735A1 (en) * 2007-06-15 2008-12-18 Los Alamos National Security Llc Anomalous change detection in imagery
WO2009008783A1 (fr) 2007-07-11 2009-01-15 Telefonaktiebolaget Lm Ericsson (Publ) Procédé et appareil pour déterminer une performance de service
US20090112932A1 (en) 2007-10-26 2009-04-30 Microsoft Corporation Visualizing key performance indicators for model-based applications
US20090238255A1 (en) * 2008-03-24 2009-09-24 Hong Kong Applied Science And Technology Research Institute Co, Ltd. Estimation of Error Propagation Probability to Improve Performance of Decision-Feedback Based Systems
US20100027432A1 (en) 2008-07-31 2010-02-04 Mazu Networks, Inc. Impact Scoring and Reducing False Positives
US20110191076A1 (en) * 2008-10-09 2011-08-04 Shunji Maeda Error detection method and system
US8254391B2 (en) * 2009-04-04 2012-08-28 Oracle International Corporation Method and system for performing blocking of messages on errors in message stream
US20100299287A1 (en) 2009-05-22 2010-11-25 Alcatel-Lucent Usa Inc. Monitoring time-varying network streams using state-space models
CN101800674A (zh) 2010-02-21 2010-08-11 浪潮通信信息系统有限公司 基于分流定向的旁路式流量检测模型
CN101854404A (zh) 2010-06-04 2010-10-06 中国科学院计算机网络信息中心 检测域名系统异常的方法和装置
US20140195184A1 (en) * 2011-08-17 2014-07-10 Hitachi, Ltd Anomaly Detection/Diagnostic Method and Anomaly Detection/Diagnostic System
US20130067106A1 (en) 2011-09-10 2013-03-14 Choudur Lakshminarayan Multi-regime detection in streaming data
US9019839B2 (en) * 2012-08-19 2015-04-28 Audiocodes Ltd. Device, system, and method of monitoring fax relay

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10548032B2 (en) * 2018-01-26 2020-01-28 Verizon Patent And Licensing Inc. Network anomaly detection and network performance status determination
US10841817B2 (en) * 2018-01-26 2020-11-17 Verizon Patent And Licensing Inc. Network anomaly detection and network performance status determination

Also Published As

Publication number Publication date
EP3075102B1 (fr) 2019-01-16
WO2015077917A1 (fr) 2015-06-04
EP3075102A1 (fr) 2016-10-05
CN105745868A (zh) 2016-07-06
US20170026250A1 (en) 2017-01-26
CN105745868B (zh) 2019-04-26
EP3075102A4 (fr) 2017-05-03

Similar Documents

Publication Publication Date Title
US10069691B2 (en) Method and apparatus for anomaly detection in a network
US10332056B2 (en) Features selection and pattern mining for KQI prediction and cause analysis
US10346756B2 (en) Machine discovery and rapid agglomeration of similar states
CN110362612B (zh) 由电子设备执行的异常数据检测方法、装置和电子设备
US10198340B2 (en) Application performance monitoring
CN108011782B (zh) 用于推送告警信息的方法和装置
JP2020017952A (ja) 警告するための方法と装置
CN109412870A (zh) 告警监控方法及平台、服务器、存储介质
CN110532152A (zh) 一种基于Kapacitor计算引擎的监控告警处理方法及系统
JP7081741B2 (ja) ネットワークデバイスの状態を判定するための方法及び装置
US12039045B2 (en) Event analysis in an electric power system
CN110661660B (zh) 告警信息根源分析方法及装置
US20140351414A1 (en) Systems And Methods For Providing Prediction-Based Dynamic Monitoring
CN113313280B (zh) 云平台的巡检方法、电子设备及非易失性存储介质
CN111541590A (zh) 链路监控方法、告警方法、系统、终端及存储介质
EP4226246A1 (fr) Analyse de cause profonde
US20200213203A1 (en) Dynamic network health monitoring using predictive functions
CN105490835A (zh) 信息监控方法和装置
CN114138601A (zh) 一种业务告警方法、装置、设备及存储介质
CN108959009A (zh) 一种服务器故障分析方法及其故障分析装置
US20230188440A1 (en) Automatic classification of correlated anomalies from a network through interpretable clustering
CN117236555A (zh) 用户能耗预测方法、设备以及存储介质
CA3120000A1 (fr) Systemes et procedes pour l'amelioration de processus pilote par des donnees
CN105607983A (zh) 数据异常监控方法和装置
CN116503802A (zh) 一种风力发电的监控方法及系统

Legal Events

Date Code Title Description
AS Assignment

Owner name: TELEFONAKTIEBOLAGET LM ERICSSON (PUBL), SWEDEN

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:HU, TRACY;LIU, QINGYAN;ZHAO, SKY;SIGNING DATES FROM 20131203 TO 20131209;REEL/FRAME:038712/0683

STCF Information on status: patent grant

Free format text: PATENTED CASE

FEPP Fee payment procedure

Free format text: MAINTENANCE FEE REMINDER MAILED (ORIGINAL EVENT CODE: REM.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

LAPS Lapse for failure to pay maintenance fees

Free format text: PATENT EXPIRED FOR FAILURE TO PAY MAINTENANCE FEES (ORIGINAL EVENT CODE: EXP.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

STCH Information on status: patent discontinuation

Free format text: PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362

FP Lapsed due to failure to pay maintenance fee

Effective date: 20220904